| View previous topic :: View next topic |
| Author |
Message |
Hu
Administrator
Joined: 06 Mar 2007
Posts: 22657
|
 Posted: Tue Dec 27, 2016 2:48 am Post subject: |
 |
|
| Fedora is a big project. Don't let a disagreeable decision from one group within the Fedora project poison your view of everything else that everyone associated with Fedora does. Sure, you might want to scrutinize more closely the decisions of the people who made that first bad decision, but that doesn't mean everyone who works on the Fedora project will make decisions you usually dislike. |
|
| Back to top |
|
 |
Zucca
Moderator
Joined: 14 Jun 2007
Posts: 3701
Location: Rasi, Finland
|
| Posted: Tue Dec 27, 2016 2:20 pm Post subject: |
|
|
| Yeah. Systemd even works in Fedora... usually. I however haven't stubled into this on Fedora, but with Arch Linux that was quite common. Boot stuck and I couldn't log into emergency shell. |
|
| Back to top |
|
|
tld
Veteran
Joined: 09 Dec 2003
Posts: 1845
|
| Posted: Tue Dec 27, 2016 2:30 pm Post subject: |
|
|
| The Doctor wrote: | | The project is still missing major functionality. |
Does Wayland have any plans on supporting anything like X11 Forwarding? Last I recall I didn't get that impression. That's something I use all the time, and which everyone seems to want to brand as a silly feature these days for some reason. |
|
| Back to top |
|
|
devilheart
l33t
Joined: 17 Mar 2005
Posts: 848
Location: Villach, Austria
|
| Posted: Tue Dec 27, 2016 2:33 pm Post subject: |
|
|
| The Doctor wrote: | That is what this solves: | package.mask wrote: | >=sys-apps/openrc-0.18
>=sys-apps/lm_sensors-3.4.0_p20160725
|
|
What happened with the more recent versions? |
|
| Back to top |
|
|
Tony0945
Watchman
Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA
|
| Posted: Tue Dec 27, 2016 3:05 pm Post subject: |
|
|
| tld wrote: | | Does Wayland have any plans on supporting anything like X11 Forwarding? Last I recall I didn't get that impression. That's something I use all the time, and which everyone seems to want to brand as a silly feature these days for some reason. |
Me too. Losing a graphical login from another box is a deal breaker. Come on! Can't X11 have a feature that Windoze has had since at least XP? |
|
| Back to top |
|
|
The Doctor
Moderator
Joined: 27 Jul 2010
Posts: 2678
|
| Posted: Wed Dec 28, 2016 3:44 am Post subject: |
|
|
| tld wrote: | | The Doctor wrote: | | The project is still missing major functionality. |
Does Wayland have any plans on supporting anything like X11 Forwarding? Last I recall I didn't get that impression. That's something I use all the time, and which everyone seems to want to brand as a silly feature these days for some reason. |
I am the wrong person to ask 
All I really know is that wayland happened because some devs didn't like they mess of code Xorg has become and want to redo it with better planning.
_________________
First things first, but not necessarily in that order.
Apologies if I take a while to respond. I'm currently working on the dematerialization circuit for my blue box. |
|
| Back to top |
|
|
The Doctor
Moderator
Joined: 27 Jul 2010
Posts: 2678
|
| Posted: Wed Dec 28, 2016 3:46 am Post subject: |
|
|
| devilheart wrote: | | The Doctor wrote: | That is what this solves: | package.mask wrote: | >=sys-apps/openrc-0.18
>=sys-apps/lm_sensors-3.4.0_p20160725
|
|
What happened with the more recent versions? |
Some sort of SNAFU. And I mean that as an abbreviation 
The newer versions won't compile without something. I don't remember what but I remember it being a deal breaker for me.
_________________
First things first, but not necessarily in that order.
Apologies if I take a while to respond. I'm currently working on the dematerialization circuit for my blue box. |
|
| Back to top |
|
|
devilheart
l33t
Joined: 17 Mar 2005
Posts: 848
Location: Villach, Austria
|
| Posted: Wed Dec 28, 2016 12:49 pm Post subject: |
|
|
| The Doctor wrote: | | devilheart wrote: | | The Doctor wrote: | That is what this solves: | package.mask wrote: | >=sys-apps/openrc-0.18
>=sys-apps/lm_sensors-3.4.0_p20160725
|
|
What happened with the more recent versions? |
Some sort of SNAFU. And I mean that as an abbreviation
The newer versions won't compile without something. I don't remember what but I remember it being a deal breaker for me. |
Mmm... it would be interesting to explore. In other threads about systemd, other users complained about recent versions of openrc, but there is no difference in the dependencies list between 0.18.4 and 0.22.4 |
|
| Back to top |
|
|
asturm
Developer
Joined: 05 Apr 2007
Posts: 9280
|
| Posted: Wed Dec 28, 2016 1:07 pm Post subject: |
|
|
| I don't see any issue with these openrc/lm_sensors versions. |
|
| Back to top |
|
|
roki942
Apprentice
Joined: 18 Apr 2005
Posts: 285
Location: Seattle
|
| Posted: Wed Dec 28, 2016 7:38 pm Post subject: |
|
|
| devilheart wrote: | | The Doctor wrote: | | devilheart wrote: | | The Doctor wrote: | That is what this solves: | package.mask wrote: | >=sys-apps/openrc-0.18
>=sys-apps/lm_sensors-3.4.0_p20160725
|
|
What happened with the more recent versions? |
Some sort of SNAFU. And I mean that as an abbreviation
The newer versions won't compile without something. I don't remember what but I remember it being a deal breaker for me. |
Mmm... it would be interesting to explore. In other threads about systemd, other users complained about recent versions of openrc, but there is no difference in the dependencies list between 0.18.4 and 0.22.4 |
When you mask openrc that way what gets installed is openrc-0.17 |
|
| Back to top |
|
|
Fitzcarraldo
Advocate
Joined: 30 Aug 2008
Posts: 2053
Location: United Kingdom
|
| Posted: Wed Dec 28, 2016 11:30 pm Post subject: |
|
|
| The Doctor wrote: | | Basically, some devs got fed up with the feature creep and messy code of xorg and decided a new project could do those things by design. We'll know if they are right in a few years. The project is still missing major functionality. |
At the risk of being labelled a Luddite, I was reminded of the xkcd Standards cartoon when I read this.
_________________
Clevo W230SS: amd64, VIDEO_CARDS="intel modesetting nvidia".
Compal NBLB2: ~amd64, xf86-video-ati. Dual boot Win 7 Pro 64-bit.
OpenRC systemd-utils[udev] elogind KDE on both.
My blog |
|
| Back to top |
|
|
berferd
Tux's lil' helper
Joined: 13 May 2004
Posts: 117
|
|
| Back to top |
|
|
tld
Veteran
Joined: 09 Dec 2003
Posts: 1845
|
| Posted: Fri Dec 30, 2016 4:34 pm Post subject: |
|
|
That silly switch from runscript to openrc-run especially pissed me off, because even masking newer versions of openrc can't prevent that from getting you burned:
I update my mythtv systems over night and had the update blow up on me. I had MythTV in my local overlay and had linked it's files directory to the portage files directory. Doing this had never been an issue before, but when the command in the init script changed, my overlay of course had an invalid digest. I wanted to scream...total silliness. |
|
| Back to top |
|
|
Naib
Watchman
Joined: 21 May 2004
Posts: 6065
Location: Removed by Neddy
|
| Posted: Fri Dec 30, 2016 6:24 pm Post subject: |
|
|
| asturm wrote: | | I don't see any issue with these openrc/lm_sensors versions. |
There was an issue, some bump in lm_sensors required a major bump in openRC (into the braindead range) thus it was simpler to remove lm_sensors and keep the openrc mask.
why did lm_sensors rely on something like that..
_________________
| Quote: | | Removed by Chiitoo |
|
|
| Back to top |
|
|
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54578
Location: 56N 3W
|
| Posted: Fri Dec 30, 2016 6:51 pm Post subject: |
|
|
Naib,
lm-sensors doesn't care. lm-sensors may need to load modules
It was made to depend on the new module list location, hence on new open-rc.
Its sensors-dectect, which can write a list of modules to be loaded, rather than lm-sensors.
If you take care of module loading for lm-sensors, the dependency on open-rc goes away.
That's how I fixed the lm-sensors ebuild anyway.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
berferd
Tux's lil' helper
Joined: 13 May 2004
Posts: 117
|
| Posted: Fri Dec 30, 2016 9:22 pm Post subject: |
|
|
| tld wrote: | | I wanted to scream...total silliness. |
Yep. Init systems can make your system unbootable after an upgrade. That ain't no joke if the system is remote and/or headless. |
|
| Back to top |
|
|
khayyam
Watchman
Joined: 07 Jun 2012
Posts: 6227
Location: Room 101
|
| Posted: Fri Dec 30, 2016 10:10 pm Post subject: |
|
|
| tld wrote: | | That silly switch from runscript to openrc-run especially pissed me off, because even masking newer versions of openrc can't prevent that from getting you burned: |
tld ... something like the following perhaps:
| /etc/local.d/openrc.stop: | #!/bin/sh
for i in $(grep --max-count=1 \
--files-with-matches \
--line-regexp '^#!/sbin/openrc-run$' \
/etc/init.d/*) ; do
sed -i -- '1s/openrc-run/runscript/' "$i" ;
done
return 0 |
edit: see below for an improved version (which excludes symbolic links).
HTH & best ... khay
Last edited by khayyam on Tue Feb 14, 2017 10:43 pm; edited 1 time in total |
|
| Back to top |
|
|
Hu
Administrator
Joined: 06 Mar 2007
Posts: 22657
|
| Posted: Sat Dec 31, 2016 5:14 pm Post subject: |
|
|
| tld wrote: | That silly switch from runscript to openrc-run especially pissed me off, because even masking newer versions of openrc can't prevent that from getting you burned:
I update my mythtv systems over night and had the update blow up on me. I had MythTV in my local overlay and had linked it's files directory to the portage files directory. Doing this had never been an issue before, but when the command in the init script changed, my overlay of course had an invalid digest. I wanted to scream...total silliness. |
The timing on your post is interesting. Last weekend, my MythTV system got rebooted for some standard upgrades and was mysteriously inaccessible over ssh afterward. I ultimately traced it down to being that I had a new openssh (which used the new #!openrc-run), but the dependencies had allowed me to install it on a system with an old openrc (that did not yet have openrc-run). As a result, /etc/init.d/sshd failed to start with Bad interpreter. For extra fun, this was not printed by the startup process, so looking at the output on the machine's console, everything looked good. The only problem was the sshd initscript never started sshd.
After breaking out a keyboard and logging in as root on the console, a manual cleanup of the style khayyam posited automating fixed it for me. As seems to be common with MythTV systems, this one is updated rarely (and usually only for targeted packages) because MythTV requires some handholding to jump versions, I don't want it broken by seemingly unrelated system changes, and I am wary of being forced into a version jump if it breaks. |
|
| Back to top |
|
|
devilheart
l33t
Joined: 17 Mar 2005
Posts: 848
Location: Villach, Austria
|
| Posted: Thu Jan 05, 2017 11:53 am Post subject: |
|
|
Thank you, I'll take a look into those threads. I have veered of Gentoo (desktop, at least) in the last years and I've lost all of these debates (especially about systemd), but now I can say there is at least one more Gentoo box inside Intel  |
|
| Back to top |
|
|
bunder
Bodhisattva
Joined: 10 Apr 2004
Posts: 5937
|
| Posted: Mon Jan 09, 2017 6:31 am Post subject: |
|
|
bug
This change was needed. As per the openrc documentation, the intent of using rc_ parameters in conf.d (eg: rc_cgroup_cpuset="cpuset.cpus 3 cpuset.mems 0") was to override rc.conf, however it was actually working in reverse. The choices were to edit every single conf.d file to your liking, force a system-wide setting in rc.conf, or ignore the rc_ parameters and use something like cgclassify. |
|
| Back to top |
|
|
khayyam
Watchman
Joined: 07 Jun 2012
Posts: 6227
Location: Room 101
|
| Posted: Mon Jan 09, 2017 8:56 am Post subject: |
|
|
| bunder wrote: |
bug. This change was needed.[...] |
bunder ... if you happen to use cgroups, then yes, but if you're setting 'rc_*_provide', 'rc_*_need', rc_*_before', or 'rc_*_config' then there is nothing to prioritise (at least I can't think of an instance in which global precedence would matter), as that was the specificity of rc.conf ... until cgroups. You're right also in that the direction of priority was in "reverse", and you would expect that local settings would override global settings, but any necessity in that regard was something introduced by the cgroups implementation.
However, all of this is beside the point, which as I read berferd, and others, is about "confidence". Take the above bug as an example: the bug goes ignored for two and a half years, then on 2016-09-05 "this [...] is the way I would prefer to fix this", and on 2016-09-06 "applies the patch ... [t]his will be in OpenRC-0.22" and 2016-09-27-openrc_0_22_updates news item is released (I assume in parallel with stabilisation of 0.22.4). Besides the long hiatus in acking, what time frame is provided for this to go through any sort of testing, and what might anyone (specifically those running "stable", or critical, systems) expect of such a change? That is not something that you can have any "confidence" in, it's willy-nilly changes to cover a problem introduced by willy-nilly changes.
best ... khay |
|
| Back to top |
|
|
bunder
Bodhisattva
Joined: 10 Apr 2004
Posts: 5937
|
| Posted: Mon Jan 09, 2017 9:18 am Post subject: |
|
|
| khayyam wrote: | | However, all of this is beside the point, which as I read berferd, and others, is about "confidence". Take the above bug as an example: the bug goes ignored for two and a half years, then on 2016-09-05 "this [...] is the way I would prefer to fix this", and on 2016-09-06 "applies the patch ... [t]his will be in OpenRC-0.22" and 2016-09-27-openrc_0_22_updates news item is released (I assume in parallel with stabilisation of 0.22.4). Besides the long hiatus in acking, what time frame is provided for this to go through any sort of testing, and what might anyone (specifically those running "stable", or critical, systems) expect of such a change? That is not something that you can have any "confidence" in, it's willy-nilly changes to cover a problem introduced by willy-nilly changes. |
For what it's worth, that bug probably would still be unfixed if it weren't for me going to freenode #openrc and reminding them about it. Sometimes people just get sidetracked or forget things. |
|
| Back to top |
|
|
augustin
Guru
Joined: 23 Feb 2015
Posts: 318
|
|
| Back to top |
|
|
khayyam
Watchman
Joined: 07 Jun 2012
Posts: 6227
Location: Room 101
|
| Posted: Tue Feb 14, 2017 10:41 pm Post subject: |
|
|
| khayyam wrote: | tld ... something like the following perhaps:
| /etc/local.d/openrc.stop: | #!/bin/sh
for i in $(grep --max-count=1 \
--files-with-matches \
--line-regexp '^#!/sbin/openrc-run$' \
/etc/init.d/*) ; do
sed -i -- '1s/openrc-run/runscript/' "$i" ;
done
return 0 |
|
... I should have taken symbolic links into account:
| /etc/local.d/openrc.stop: | #!/bin/sh
for i in $(find /etc/init.d/ -type f -a -exec \
grep \
--max-count=1 \
--files-with-matches \
--line-regexp '^#!/sbin/openrc-run$' '{}' \;) ; do
sed -i -- '1s/openrc-run/runscript/' "$i" ;
done |
best ... khay |
|
| Back to top |
|
|
tld
Veteran
Joined: 09 Dec 2003
Posts: 1845
|
| Posted: Wed Feb 15, 2017 3:49 pm Post subject: |
|
|
| Thanks khay! |
|
| Back to top |
|
|
|
Gentoo Chat
